U.S. patent number 11,249,159 [Application Number 16/899,357] was granted by the patent office on 2022-02-15 for systems and methods for enhancement of resolution for strategically acquired gradient echo (stage) imaging.
This patent grant is currently assigned to SPINTECH, INC.. The grantee listed for this patent is SPINTECH, INC.. Invention is credited to Yongsheng Chen, E. Mark Haacke.
United States Patent |
11,249,159 |
Haacke , et al. |
February 15, 2022 |
Systems and methods for enhancement of resolution for strategically
acquired gradient echo (stage) imaging
Abstract
Systems and methods for high-resolution STAGE imaging can
include acquisition of relatively low-resolution k-space datasets
with two separate multi-echo GRE sequences. The multi-echo GRE
sequences can correspond to separate and distinct flip angles.
Various techniques for combining the low-resolution k-space
datasets to generate a relatively high-resolution k-space are
described. These techniques can involve combining low-resolution
k-space datasets associated with various echo types. The STAGE
imaging approaches described herein allow for rapid imaging,
enhanced image resolution with relatively small or no increase in
MR data acquisition time.
Inventors: |
Haacke; E. Mark (Detroit,
MI), Chen; Yongsheng (Detroit, MI) |
Applicant: |
Name |
City |
State |
Country |
Type |
SPINTECH, INC. |
Bingham Farms |
MI |
US |
|
|
Assignee: |
SPINTECH, INC. (Bingham Farms,
MI)
|
Family
ID: |
1000006115173 |
Appl.
No.: |
16/899,357 |
Filed: |
June 11, 2020 |
Prior Publication Data
|
|
|
|
Document
Identifier |
Publication Date |
|
US 20210389401 A1 |
Dec 16, 2021 |
|
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
A61B
5/055 (20130101); G01R 33/5608 (20130101); G01R
33/5615 (20130101); G01R 33/4826 (20130101) |
Current International
Class: |
G01V
3/00 (20060101); G01R 33/561 (20060101); G01R
33/56 (20060101); A61B 5/055 (20060101); G01R
33/48 (20060101) |
Field of
Search: |
;324/309 |
References Cited
[Referenced By]
U.S. Patent Documents
Other References
Chen Y, Liu S, Wang Y, Kang Y, Haacke EM. STrategically Acquired
Gradient Echo (STAGE) imaging, part I: Creating enhanced T1
contrast and standardized susceptibility weighted imaging and
quantitative susceptibility mapping. Magn. Reson. Imaging
2018;46:130-139. doi: 10.1016/j.mri.2017.10.005. cited by applicant
.
Do, Won-Joon, Seung Hong Choi, and Sung-Hong Park. "Simultaneous
Variable-Slab Dual-Echo TOF MR Angiography and
Susceptibility-Weighted Imaging." IEEE transactions on medical
imaging 37.7 (2018): 1632-1640. cited by applicant .
Wang Y, Chen Y, Wu D, Wang Y, Sethi SK, Yang G, Xie H, Xia S,
Haacke EM. STrategically Acquired Gradient Echo (STAGE) imaging,
part II: Correcting for RF inhomogeneities in estimating T1 and
proton density. Magn Reson Imaging 2018;46:140-50. doi:
10.1016/j.mri.2017.10.006. cited by applicant.
|
Primary Examiner: Lindsay, Jr.; Walter L
Assistant Examiner: Wenderoth; Frederick
Attorney, Agent or Firm: Foley & Lardner LLP
Claims
What is claimed is:
1. A magnetic resonance imaging (MRI) system comprising: an MRI
scanner configured to: acquire, for a first flip angle, a first
magnetic resonance (MR) k-space dataset and a second MR k-space
dataset by scanning an anatomical region of interest with at least
two echo times, the first MR k-space dataset acquired at a first
TE1 echo time and the second MR k-space dataset acquired at a first
TE2 echo time; acquire, for a second flip angle different from the
first flip angle, a third MR k-space dataset and a fourth MR
k-space dataset by scanning the anatomical region of interest with
the at least two echo times, the third MR k-space dataset acquired
at a second TE1 echo time different from the first TE1 echo time
and the fourth MR k-space dataset acquired at a second TE2 echo
time different from the first TE2 echo time; at least one
processor; and a memory, with computer code instructions stored
thereon, the computer code instructions, when executed by the at
least one processor, cause the at least one processor to: generate
a fifth MR k-space dataset by combining the fourth MR k-space
dataset with either (i) the second MR k-space dataset or (ii) a
central extrapolation of the third MR k-space dataset; and
reconstruct an MR image of the anatomical region of interest using
the fifth MR k-space dataset.
2. The MRI system of claim 1, wherein the at least one processor is
further configured to reconstruct a susceptibility-weighted image
using the MR image of the anatomical region of interest.
3. The MRI system of claim 1, wherein the at least one processor is
further configured to reconstruct a quantitative susceptibility
mapping (QSM) image using the MR image of the anatomical region of
interest.
4. The MRI system of claim 1, wherein combining the second MR
k-space dataset and the fourth MR k-space dataset includes: using
the second MR k-space dataset to generate a central portion of the
fifth MR k-space dataset; and using the fourth MR k-space dataset
to generate two opposite outer portions of the fifth MR k-space
dataset.
5. The MRI system of claim 4, wherein the central portion of the
modified third echo MR k-space dataset (i) partially overlaps with
a first outer portion of the two opposite outer portions of the
fifth MR k-space dataset along a first overlap region, and (ii)
partially overlaps with a second outer portion of the two opposite
outer portions of the fifth MR k-space dataset along a second
overlap region.
6. The MRI system of claim 1, wherein the computer code
instructions, when executed by the at least one processor, cause
the at least one processor to: generate a second MR image by using
an inverse Fourier transform of the first MR k-space dataset;
generate a third MR image by using an inverse Fourier transform of
the second MR k-space data set; complex divide the third MR image
by the second MR image to obtain a phase difference image and a T2*
weighting factor; generate a fourth MR image by using an inverse
Fourier transform of the third MR k-space data set; adjust, using
the phase difference image and the T2* weighting factor, the fourth
MR image to generate a fifth MR image; Fourier transform the fifth
MR image to obtain the central k-space extrapolation of the third
MR k-space dataset; generate the fifth MR k-space dataset by
combining the central extrapolation of the third MR k-space dataset
and the fourth MR data k-space dataset, the central extrapolation
of the third MR k-space dataset to generate a central portion of
the fifth MR k-space dataset and the fourth MR k-space dataset used
to generate two opposite outer portions of the fifth MR k-space
dataset; and apply inverse Fourier transform to the fifth MR
k-space dataset to reconstruct the MR image of the anatomical
region of interest.
7. The MRI system of claim 6, wherein the central extrapolation of
the third MR k-space dataset (i) partially overlaps with a first
outer portion of the two opposite outer portions of the fifth MR
k-space dataset along a first overlap region, and (ii) partially
overlaps with a second outer portion of the two opposite outer
portions of the fifth MR k-space dataset along a second overlap
region.
8. The MRI system of claim 1, wherein combining the second MR
k-space dataset and the fourth MR k-space dataset includes: using
the second MR k-space dataset to generate a first side portion of
the fifth MR k-space dataset; and using the fourth MR k-space
dataset to generate a second side portion of the fifth MR k-space
dataset, the first side portion opposite to the second side
portion.
9. The MRI system of claim 6, wherein the first side portion of the
fifth MR k-space dataset partially overlaps with the second side
portion of the fifth MR k-space dataset along an overlap
region.
10. The MRI system of claim 9, wherein in generating the fifth MR
k-space dataset, the at least one processor is configured to:
generate a second MR image by using an inverse Fourier transform of
the second MR k-space data set; generate a third MR image by using
an inverse Fourier transform of the fourth MR k-space data set;
compare phase information of the second MR image to phase
information of the third MR image; adjust, based on the comparison,
the phase information of the third MR image so that the adjusted
phase information of the third MR image is equal to the phase
information of the second MR image; generate a sixth MR k-space
dataset by applying a Fourier transform to the third MR image with
the adjusted phase information; and generate the fifth MR k-space
dataset by combining the second MR k-space dataset and the sixth MR
k-space dataset within the overlap region.
11. The MRI system of claim 1, wherein the first flip angle is 6
degrees and the second flip angle is 24 degrees.
12. A method for magnetic resonance imaging (MRI), comprising:
acquiring, by an MRI scanner, for a first flip angle, a first
magnetic resonance (MR) k-space dataset and a second MR k-space
dataset by scanning an anatomical region of interest with at least
two echo times, the first MR k-space dataset acquired at a first
TE1 echo time and the second MR k-space dataset acquired at a first
TE2 echo time; acquiring, by the MRI scanner, for a second flip
angle different from the first flip angle, a third MR k-space
dataset and a fourth MR k-space dataset by scanning the anatomical
region of interest with the at least two echo times, the third MR
k-space dataset acquired at a second TE1 echo time different from
the first TE1 echo time and the fourth MR k-space dataset acquired
at a second TE2 echo time different from the first TE2 echo time;
generating, by the MRI scanner, a fifth MR k-space dataset by
combining the fourth MR k-space dataset and either (i) the second
MR k-space dataset or (ii) a central extrapolation of the third MR
k-space dataset; and reconstructing, by the MRI scanner, an MR
image of the anatomical region of interest using the fifth MR
k-space dataset.
13. The method of claim 12, further comprising reconstructing a
susceptibility-weighted image or a quantitative susceptibility
mapping (QSM) image using the MR image of the anatomical region of
interest.
14. The method of claim 12, comprising: generating a second MR
image by using an inverse Fourier transform of the first MR k-space
dataset; generating a third MR image by using an inverse Fourier
transform of the second MR k-space data set; complex dividing the
third MR image by the second MR image to obtain a phase difference
image and a T2* weighting factor; generating a fourth MR image by
using an inverse Fourier transform of the third MR k-space data
set; adjusting, using the phase difference image and the T2*
weighting factor, the fourth MR image to generate a fifth MR image;
Fourier transforming the fifth MR image to obtain the central
extrapolation of the third MR k-space dataset; generate the fifth
MR k-space dataset by combining the central extrapolation of the
third MR k-space dataset and the fourth MR data k-space dataset,
the central extrapolation of the third MR k-space dataset to used
generate a central portion of the fifth MR k-space dataset and the
fourth MR k-space dataset used to generate two opposite outer
portions of the fifth MR k-space dataset; and applying inverse
Fourier transform to the fifth MR k-space dataset to reconstruct
the MR image of the anatomical region of interest.
15. The method of claim 14, wherein the central extrapolation of
the third MR k-space dataset (i) partially overlaps with a first
outer portion of the two opposite outer portions of the fifth MR
k-space dataset along a first overlap region, and (ii) partially
overlaps with a second outer portion of the two opposite outer
portions of the fifth MR k-space dataset along a second overlap
region.
16. The method of claim 12, wherein combining the second MR k-space
dataset and the fourth MR k-space dataset includes: using the
second MR k-space dataset to generate a central portion of the
fifth MR k-space dataset; and using the fourth MR k-space dataset
to generate two opposite outer portions of fifth MR k-space
dataset.
17. The method of claim 16, wherein the central portion of the
fifth MR k-space dataset (i) partially overlaps with a first outer
portion of the two opposite outer portions of the fifth MR k-space
dataset along a first overlap region, and (ii) partially overlaps
with a second outer portion of the two opposite outer portions of
the fifth MR k-space dataset along a second overlap region.
18. The method of claim 12, wherein combining the second MR k-space
dataset and the fourth MR k-space dataset includes: using the
second MR k-space dataset to generate a first side portion of the
fifth MR k-space dataset; and using the fourth MR k-space dataset
to generate a second side portion of the fifth MR k-space dataset,
the first side portion opposite to the second side portion.
19. The method of claim 17, wherein the first side portion of the
fifth MR k-space dataset partially overlaps with the second side
portion of the fifth MR k-space dataset along an overlap
region.
20. The method of claim 18, wherein generating the fifth MR k-space
dataset includes: generating a second MR image by using a Fourier
transform of the second MR k-space data set; generating a third MR
image by using a Fourier transform of the fourth MR k-space data
set; comparing phase information of the second MR image to phase
information of the third MR image; adjusting, based on the
comparison, the phase information of the third MR image so that the
adjusted phase information of the third MR image is equal to the
phase information of the second MR image; generating a sixth MR
k-space dataset by applying an inverse Fourier transform to the
third MR image with the adjusted phase information; and generating
the fifth MR k-space dataset by combining the second MR k-space
dataset and the sixth MR k-space dataset within the overlap
region.
21. The method of claim 12, wherein the first flip angle is 6
degrees and the second flip angle is 24 degrees.
22. A non-transitory computer-readable medium comprising computer
code instructions stored thereon, the computer code instructions
when executed by at least one processor cause the at least one
processor to: cause a magnetic resonance imaging (MRI) scanner to
acquire, for a first flip angle, a first magnetic resonance (MR)
k-space dataset and a second 1\4R k-space dataset by scanning an
anatomical region of interest with at least two echo times, the
first MR k-space dataset acquired at a first TE1 echo time and the
second MR k-space dataset acquired at a first TE2 echo time; cause
the MRI scanner to acquire, for a second flip angle different from
the first flip angle, a third MR k-space dataset and a fourth MR
k-space dataset by scanning the anatomical region of interest with
the at least two echoes times, the third MR k-space dataset
acquired at a second TE1 echo time different from the first TE1
echo time and the fourth MR k-space dataset acquired at a second
TE2 echo time different from the first TE2 echo time; generate a
fifth MR k-space dataset by combining the fourth MR k-space dataset
and either (i) the second MR k-space dataset or (ii) a central
extrapolation of the third MR k-space dataset; and reconstruct an
MR image of the anatomical region of interest using the fifth MR
k-space dataset.
Description
BACKGROUND OF THE DISCLOSURE
The present disclosure relates generally to the field of magnetic
resonance imaging (MRI). Specifically, this disclosure relates to
methods and systems for improving the resolution of strategically
acquired gradient echo (STAGE) imaging that involves the use of
more than one flip angle.
Magnetic resonance imaging (MRI) is an imaging modality that uses
magnetic fields to reconstruct an image representing part or all of
the scanned object or person of interest. An MRI scanner includes a
magnet for generating a strong static magnetic field, such as a
magnetic field in the range of 0.05 Tesla (T) to 20 T, and radio
frequency (RF) transceivers for transmitting and/or receiving RF
signals. When a body is placed in the generated static magnetic
field, the hydrogen protons within the body align to the magnetic
field. An RF pulse is applied in the form of an oscillating B1
field to tip the spins so that there is a bulk magnetization
remaining in the transverse field. When the RF pulse is turned off,
the hydrogen protons gradually realign with the static magnetic
field. The RF receiver coils detect the precessing magnetization
and from it create a measurable current. At predefined time points,
referred to as the sampling time, echo time (TE), or gradient echo
time, data are collected and sampled and used to reconstruct an
image of the scanned body or a part thereof.
SUMMARY OF THE DISCLOSURE
According to at least one aspect, a magnetic resonance imaging
(MRI) system can include an MRI scanner, at least one processor,
and a memory, with computer code instructions stored thereon. The
MRI scanner can acquire, for a first flip angle, a first magnetic
resonance (MR) k-space dataset and a second MR k-space dataset by
scanning an anatomical region of interest with at least two echo
times. The MRI scanner can acquire the first MR k-space dataset at
a first TE1 echo time and acquire the second MR k-space dataset at
a first TE2 echo time. The MRI scanner can acquire, for a second
flip angle, a third MR k-space dataset and a fourth MR k-space
dataset by scanning the anatomical region of interest with the at
least two echo times. The MRI scanner can acquire the third MR
k-space dataset at a second TE1 echo time and acquire the fourth MR
k-space dataset at a second TE2 echo time. The computer code
instructions, when executed by the at least one processor, can
cause the at least one processor to generate a fifth MR k-space
dataset by combining the fourth MR k-space dataset with either (i)
the second MR k-space dataset or (ii) a central extrapolation of
the third MR k-space dataset. The at least one processor can
reconstruct an MR image of the anatomical region of interest using
the fifth MR k-space dataset.
The at least one processor can be configured to reconstruct a
susceptibility-weighted image using the MR image of the anatomical
region. The at least one processor can be configured to reconstruct
a quantitative susceptibility mapping (QSM) image using the MR
image of the anatomical region. In reconstructing the MR image of
the anatomical region of interest, the at least one processor can
be configured to compute an inverse Fourier transform of the fifth
MR k-space dataset. In some implementations, the first flip angle
can be equal to 6 degrees and the second flip angle can be equal to
24 degrees.
The at least one processor can, in combining the second MR k-space
dataset and the fourth MR k-space dataset, use the second MR
k-space dataset to generate a central portion of the fifth MR
k-space dataset and use the fourth MR k-space dataset to generate
two opposite outer portions of the fifth MR k-space dataset. The
central portion of the fifth MR k-space dataset can (i) partially
overlap with a first outer portion of the two opposite outer
portions of the fifth MR k-space dataset along a first overlap
region, and (ii) partially overlap with a second outer portion of
the two opposite outer portions of the fifth MR k-space dataset
along a second overlap region.
The at least one processor can generate a second MR image by using
an inverse Fourier transform of the first MR k-space dataset, and
generate a third MR image by using an inverse Fourier transform of
the second MR k-space data set. The at least one processor can
complex divide the third MR image by the second MR image to obtain
a phase difference image and a T2* weighting factor. The at least
one processor can generate a fourth MR image by using an inverse
Fourier transform of the third MR k-space data set. The at least
one processor can adjust, using the phase difference image and the
T2* weighting factor, the fourth MR image to generate a fifth MR
image. The at least one processor can Fourier transform the fifth
MR image to obtain the central k-space extrapolation of the third
MR k-space dataset. The at least one processor can generate the
fifth MR k-space dataset by combining the central extrapolation of
the third MR k-space dataset and the fourth MR data k-space
dataset. The central extrapolation of the third MR k-space dataset
can be used to generate a central portion of the fifth MR k-space
dataset and the fourth MR k-space dataset can be used to generate
two opposite outer portions of the fifth MR k-space dataset. The at
least one processor can apply inverse Fourier transform to the
fifth MR k-space dataset to reconstruct the MR image of the
anatomical region of interest. The central extrapolation of the
third MR k-space dataset can (i) partially overlap with a first
outer portion of the two opposite outer portions of the fifth MR
k-space dataset along a first overlap region, and (ii) partially
overlap with a second outer portion of the two opposite outer
portions of the fifth MR k-space dataset along a second overlap
region.
In combining the second MR k-space dataset and the fourth MR
k-space dataset, the at least one processor can use the second MR
k-space dataset to generate a first side portion of the fifth MR
k-space dataset and use the fourth MR k-space dataset to generate a
second side portion of the fifth MR k-space dataset. The first side
portion can be to opposite to the second side portion. The first
side portion of the fifth MR k-space dataset can partially overlap
with the second side portion of the fifth MR k-space dataset along
an overlap region.
In generating the fifth MR k-space dataset, the at least one
processor can be configured to generate a second MR image by using
an inverse Fourier transform of the second MR k-space data set and
generate a third MR image by using an inverse Fourier transform of
the third MR k-space data set. The at least one processor can
compare phase information of the second MR image to phase
information of the third MR image. The at least one processor can
adjust, based on the comparison, the phase information of the third
MR image so that the adjusted phase information of the third MR
image is equal to the phase information of the second MR image. The
at least one processor can generate a sixth MR k-space dataset by
applying a Fourier transform to the third MR image with the
adjusted phase information. The at least one processor can generate
the fifth MR k-space dataset by combining the second MR k-space
dataset and the sixth MR k-space dataset within the overlap
region.
According to at least one aspect, a method for magnetic resonance
imaging (MRI) can include an MRI scanner acquiring, for a first
flip angle, a first magnetic resonance (MR) k-space dataset and a
second MR k-space dataset by scanning an anatomical region of
interest with at least two echo times. The MRI scanner can acquire
the first MR k-space dataset at a first TE1 echo time and the
second MR k-space dataset at a first TE2 echo time. The method can
include the MRI scanner acquiring, for a second flip angle, a third
MR k-space dataset and a fourth MR k-space dataset by scanning the
anatomical region of interest with the at least two echo times. The
MRI scanner can acquire the third MR k-space dataset at a second
TE1 echo time and acquire the fourth MR k-space dataset at a second
TE2 echo time. The method can include the MRI scanner generating a
fifth MR k-space dataset by combining the fourth MR k-space dataset
with either (i) the second MR k-space dataset or (ii) a central
extrapolation of the third MR k-space dataset. The method can
include the MRI scanner reconstructing an MR image of the
anatomical region of interest using the fifth MR k-space
dataset.
The method can include reconstructing a susceptibility-weighted
image using the MR image of the anatomical region. The method can
also include reconstructing a quantitative susceptibility mapping
(QSM) image using the MR image of the anatomical region.
Reconstructing the MR image of the anatomical region of interest
can include computing an inverse Fourier transform of the fifth MR
k-space dataset. The first flip angle can be equal to 6 degrees and
the second flip angle can be equal to 24 degrees.
Combining the second MR k-space dataset and the fourth MR k-space
dataset can include using the second MR k-space dataset to generate
a central portion of the fifth MR k-space dataset, and using the
fourth MR k-space dataset to generate two opposite outer portions
of fifth MR k-space dataset. The central portion of the fifth MR
k-space dataset (i) can partially overlap with a first outer
portion of the two opposite outer portions of the fifth MR k-space
dataset along a first overlap region, and (ii) can partially
overlap with a second outer portion of the two opposite outer
portions of the fifth MR k-space dataset along a second overlap
region.
The method can include generating a second MR image by using an
inverse Fourier transform of the first MR k-space dataset, and
generating a third MR image by using an inverse Fourier transform
of the second MR k-space data set. The method can include complex
dividing the third MR image by the second MR image to obtain a
phase difference image and a T2* weighting factor. The method can
include generating a fourth MR image by using an inverse Fourier
transform of the third MR k-space data set. The method can include
adjusting, using the phase difference image and the T2* weighting
factor, the fourth MR image to generate a fifth MR image. The
method can include Fourier transforming the fifth MR image to
obtain the central k-space extrapolation of the third MR k-space
dataset. The method can include generating the fifth MR k-space
dataset by combining the central extrapolation of the third MR
k-space dataset and the fourth MR data k-space dataset. The central
extrapolation of the third MR k-space dataset can be used to
generate a central portion of the fifth MR k-space dataset and the
fourth MR k-space dataset can be used to generate two opposite
outer portions of the fifth MR k-space dataset. The method can
include applying inverse Fourier transform to the fifth MR k-space
dataset to reconstruct the MR image of the anatomical region of
interest. The central extrapolation of the third MR k-space dataset
can (i) partially overlap with a first outer portion of the two
opposite outer portions of the fifth MR k-space dataset along a
first overlap region, and (ii) partially overlap with a second
outer portion of the two opposite outer portions of the fifth MR
k-space dataset along a second overlap region.
Combining the second MR k-space dataset and the fourth MR k-space
dataset can include using the second MR k-space dataset to generate
a first side portion of the fifth MR k-space dataset, and using the
fourth MR k-space dataset to generate a second side portion of the
fifth MR k-space dataset. The first side portion can be opposite to
the second side portion. The first side portion of the fifth MR
k-space dataset can partially overlap with the second side portion
of the fifth MR k-space dataset along an overlap region.
Generating the fifth MR k-space dataset can include generating a
second MR image by using an inverse Fourier transform of the second
MR k-space dataset and generating a third MR image by using an
inverse Fourier transform of the fourth MR k-space data set. The
method can include comparing phase information of the second MR
image to phase information of the third MR image. The method can
include adjusting, based on the comparison, the phase information
of the third MR image so that the adjusted phase information of the
third MR image is equal to the phase information of the second MR
image. The method can include generating a sixth MR k-space dataset
by applying a Fourier transform to the third MR image with the
adjusted phase information. The method can include generating the
fifth MR k-space dataset by combining the second MR k-space dataset
and the sixth MR k-space dataset within the overlap regions.
According to at least one aspect, a computer-readable medium can
include computer code instructions stored thereon. The computer
code instructions, when executed by at least one processor, can
cause the at least one processor to cause a magnetic resonance
imaging (MRI scanner to acquire, for a first flip angle, a first
magnetic resonance (MR k-space dataset and a second MR k-space
dataset by scanning an anatomical region of interest with at least
two echo times. The first MR k-space dataset can be acquired at a
first TE1 echo time and the second MR k-space dataset can be
acquired at a first TE2 echo time. The at least one processor can
cause the MRI scanner to acquire, for a second flip angle, a third
MR k-space dataset and a fourth MR k-space dataset by scanning the
anatomical region of interest with the at least two echoes times.
The third MR k-space dataset can be acquired at a second TE1 echo
time and the fourth MR k-space dataset can be acquired at a second
TE2 echo time. The at least one processor can generate a fifth MR
k-space dataset by combining the fourth MR k-space dataset with
either (i) the second MR k-space dataset or (ii) a central
extrapolation of the third MR k-space dataset. The at least one
processor can reconstruct an MR image of the anatomical region of
interest using the fifth MR k-space dataset.
According to at least one aspect, a magnetic resonance imaging
(MRI) system can include an MRI scanner, at least one processor,
and a memory, with computer code instructions stored thereon. The
MRI scanner can acquire, for a first flip angle, a first magnetic
resonance (MR) k-space dataset of an anatomical region of interest.
The MRI scanner can acquire, for a second flip angle, a second MR
k-space dataset of the anatomical region of interest. The computer
code instructions, when executed by the at least one processor, can
cause the at least one processor to generate a third MR k-space
dataset by combining the first MR k-space dataset and the second MR
k-space dataset. The at least one processor can reconstruct an MR
image of the anatomical region of interest using the third MR
k-space dataset.
According to at least one aspect, a method for magnetic resonance
imaging (MRI) can include an MRI scanner acquiring, for a first
flip angle, a first magnetic resonance (MR) k-space dataset of an
anatomical region of interest. The method can include the MRI
scanner acquiring, for a second flip angle, a second MR k-space
dataset of the anatomical region of interest. The method can
include the MRI scanner generating a third MR k-space dataset by
combining the first MR k-space dataset and the second MR k-space
dataset. The method can include the MRI scanner reconstructing an
MR image of the anatomical region of interest using the third MR
k-space dataset.
According to at least one aspect, a computer-readable medium can
include computer code instructions stored thereon. The computer
code instructions, when executed by at least one processor, can
cause the at least one processor to cause an MRI scanner to
acquire, for a first flip angle, a first magnetic resonance (MR)
k-space dataset of an anatomical region of interest. The at least
one processor can cause the MRI scanner to acquire, for a second
flip angle, a second MR k-space dataset of the anatomical region of
interest. The computer code instructions, when executed by the at
least one processor, can cause the at least one processor to
generate a third MR k-space dataset by combining the first MR
k-space dataset and the second MR k-space dataset. The at least one
processor can reconstruct an MR image of the anatomical region of
interest using the third MR k-space dataset.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a block diagram illustrating a magnetic resonance imaging
(MRI) system, according to inventive concepts of this
disclosure;
FIG. 2 is a flowchart illustrating a method of strategically
acquired gradient echo (STAGE) imaging, according to inventive
concepts of this disclosure;
FIG. 3 is a flowchart illustrating another method of strategically
acquired gradient echo (STAGE) imaging, according to inventive
concepts of this disclosure;
FIGS. 4A and 4B show block diagrams depicting example approaches
for acquiring and combining MR data corresponding to multiple flip
angles, according to inventive concepts of this disclosure;
FIG. 5 shows a block diagram depicting another example approach for
acquiring and combining MR data corresponding to multiple flip
angles, according to inventive concepts of this disclosure;
FIG. 6 is a flowchart illustrating yet another example approach of
strategically acquired gradient echo (STAGE) imaging, according to
inventive concepts of this disclosure;
FIG. 7 shows various MR brain images corresponding to distinct flip
angles for TE1 data are shown;
FIG. 8 shows a block diagram depicting yet another example approach
for acquiring and combining MR data corresponding to multiple flip
angles, according to inventive concepts of this disclosure;
FIG. 9 shows STAGE imaging simulation results for various types of
MR images; and
FIG. 10 shows STAGE imaging simulation results illustrating the
advantages of k-space merging of the two second echoes from each of
the different flip angles.
DETAILED DESCRIPTION
Strategically acquired gradient echo (STAGE) imaging is a rapid
multi-contrast imaging method used to collect clinical data very
quickly usually in less than 5 minutes, for a typical
implementation at 3 T, depending on the desired resolution. In some
implementations, data acquisition can take more than 5 minutes if a
relatively high resolution is sought. A magnetic resonance (MR)
imaging system can use two different flip angles to create, for
example, proton spin density weighted (PDW) images and T1W images.
The MR data can be collected with a short repetition time (TR) on
the order of 25 ms and with flip angles of 6.degree. (e.g., for PDW
images) and 24.degree. (e.g., for T1W images) at 3 T. It should be
noted that other TR and/or other flip angles can be used.
Collecting MR data at separate sequences corresponding to distinct
flip angles can lead to MR data with different contrast. The
difference in contrast usually means that information from the two
sequences (or scans) would result in poor image quality or
undesired image artifacts if combined or merged together. However,
here, various embodiments of STAGE imaging that involve combining
k-space datasets corresponding to distinct flip angles are
described and illustrated for both phase data and for some
combinations of magnitude data as well.
An example application, the concept of data sharing (or merging)
across dual echo (or multi-echo) sequences corresponding to
distinct flip angles can be used to create susceptibility weighted
imaging (SWI) data. SWI has played a role in more than 1000 studies
since its inception. The use of SWI relies to a large degree on
collecting the data with the right resolution, usually a higher
resolution than that collected in the existing rapid STAGE
protocol. Such low resolution does not reveal the small veins well
and, hence, does not satisfy the needs of most clinical studies
with SWI or for quantitative susceptibility mapping (QSM). The
STAGE imaging approaches described herein alleviates this problem
by doubling (or increasing) the in-plane resolution of the second
echo (or later echoes) data to achieve higher quality SWI and QSM
images. While one would expect the resulting increase in resolution
would lead to a proportional increase in data acquisition time, the
STAGE imaging approaches described herein have little or no impact
on the time duration of the MR data acquisition. Both SWI and QSM
are important for studying calcifications, asymmetrically prominent
cortical veins (for stroke), damaged veins in traumatic brain
injury (TBI), iron measurements (for multiple sclerosis and
Parkinson's disease) and oxygen saturation measurements (for
stroke).
In the current disclosure, several k-space sharing (or merging)
techniques provide relatively high resolution SWI and QSM while
maintaining rapid scan time. These techniques can include acquiring
the high frequency and low frequency k-space datasets in,
respectively, a first dual-echo (or multi-echo) gradient echo (GRE)
sequence corresponding to a first flip angle (e.g., 6 degrees) and
a second dual-echo (or multi-echo) GRE sequence corresponding to a
second flip angle (e.g., 24 degrees). The STAGE imaging approaches
described herein can be referred to as high resolution SWI STAGE
(HR-SWI-STAGE). The STAGE imaging approaches described herein also
address the case where MR data acquired at distinct dual-echo (or
multi-echo) GRE sequences are associated with different
contrast.
The STAGE imaging techniques described herein involve data
acquisition for two dual-echo (or multi-echo) GRE sequences
corresponding to two different flip angles. The total acquisition
time can be around 5 minutes, for example, with the parameters
listed in Table 1 below. To assess the performance of the STAGE
imaging approaches described herein, a fully sampled
high-resolution data set was collected. Subsets of the collected
data are used as low-resolution k-space datasets. The
low-resolution k-space datasets represent only portions of the
corresponding k-spaces. Combining the different k-space coverages
as described herein allows for creating a high resolution k-space
coverage that can then be inverse Fourier transformed to create a
high-resolution MR image. The STAGE imaging techniques described
herein result in little to no increase in data acquisition
time.
In the current disclosure, methods and systems for STAGE imaging
with improved image resolution are described. The STAGE imaging
techniques described herein allow for rapid imaging, enhanced image
resolution and/or improved SNR by using acquired k-space datasets
corresponding to distinct flip angles.
TABLE-US-00001 TABLE 1 Example STAGE imaging parameters for 3T.
Axial Dual-echo GRE Axial Dual-echo GRE Read .times. Phase FOV (mm)
256 .times. 192 256 .times. 192 Scanning matrix 384 .times. 144 384
.times. 144 Voxel size (mm.sup.3) 0.67 .times. 1.33 .times. 2.0
0.67 .times. 1.33 .times. 2.0 Number of slices 64 64 Slice
oversampling 12.5% 12.5% TR (ms) 25 25 lEs (ms) 7.5, 17.5 7.5, 17.5
FA (degree) 6 24 Sampling bandwidth (Hz/ 240 240 pixel) Fully flow
compensation Yes Yes Acc. Factor (GRAPPA) 2 2 TA (min:sec) 2:29
2:29
FIG. 1 is a block diagram illustrating an MRI system 100, according
to inventive concepts of this disclosure. In brief overview, the
MRI system 100 can include an MRI scanner 102, a processor 104, a
memory 106, and a display device 108. The processor 104 can be
communicatively coupled to the MRI scanner 102, the memory 106 and
the display device 108. In some implementations, the processor 104,
the memory 106, the display device 108 or a combination thereof can
be components of the MRI scanner 102. The MRI scanner 102 can
include a magnet (not shown in FIG. 1) for generating a relatively
strong static magnetic field, such as a magnetic field in the range
of 0.05 Tesla (T) to 20 T. The magnet can have a cylindrical shape
forming a cavity 110 designed to receive a patient or other
subject. The MRI scanner 102 can include a sliding table 110. The
patient can lie down on the sliding table 112, and the position of
the sliding table 112 can be adjusted such that an anatomical
region of interest of the patient, e.g., the patient's head or
chest, falls within the cavity 110 and is subjected to the magnetic
field generated by the magnet.
The MRI scanner 102 can include a plurality of radio frequency (RF)
coils (not shown in FIG. 1) for transmitting and/or receiving RF
signals. The RF coils can include transmit RF coils and receive RF
coils. The RF transmit coils can emit RF pulses to excite the
anatomical region of interest of the patient, according to an MRI
pulse sequence. The receive RF coils can record MRI signals
generated by the anatomical region of interest following completion
of the RF transmit pulse. The RF coils may include RF transceivers
capable of alternately transmitting and receiving RF signals. The
RF coils can include gradient coils designed to induce gradients,
or distortions, in the main magnetic field generated by the magnet
in a predictable or a predefined way to make spatial separation of
the different spatial components of the object uniquely.
Specifically, the gradient coils can include frequency-encoding
gradients and phase-encoding gradients.
The imaging system 100 can include one or more processors 104. The
one or more processors 104 can include a processor integrated
within the MRI scanner 102, a processor of a computing device
communicatively coupled to the MRI scanner 102, or a combination
thereof. The memory 106 can include a memory component of the MRI
scanner 102, a memory component of a computing device
communicatively coupled to the MRI scanner 102, or a combination
thereof. The memory 106 can include computer executable
instructions, which when executed by the one or more processors
104, can cause the one or more processors 104 to perform methods
for STAGE imaging described herein. The memory 106 can store MRI
data acquired by the MRI scanner 102, and the processor(s) 104 can
access such data from the memory 106. The memory 106 can receive
and store images generated by the processor(s) 104 based on the MRI
data acquired by the scanner 102.
The display device 108 can include a cathode ray tube (CRT)
display, a light emitting diode (LED) display, an organic light
emitting diode (OLED) display, a plasma display panel (PDP), a
liquid crystal display (LCD), or other display known to a person of
ordinary skill in the art. The display device 108 may be a
stand-alone device or a display of a computing device (e.g., a
desktop, laptop, or tablet) communicatively coupled to the MRI
scanner 102. The display device 108 can include a touch screen. The
display device 108 can receive image data from the processor 104 or
the memory 106 and display the received image data. For example,
upon reconstructing MRI images based on data acquired by the MRI
scanner 102, the processor 104 can provide the reconstructed images
for display on the display device 108.
FIG. 2 is a flowchart illustrating a method 200 of strategically
acquired gradient echo (STAGE), according to inventive concepts of
this disclosure. In brief overview, the method 200 can include
acquiring a first MR k-space dataset of an anatomical region of
interest corresponding to a first flip angle (STEP 202), and
acquiring a second MR k-space dataset of the anatomical region of
interest corresponding to a second flip angle (STEP 204). The
method 200 can include generating a third MR k-space dataset by
combining the first MR k-space dataset and the second MR k-space
dataset (STEP 206). The method 200 can include constructing an MR
image of the anatomical region of interest using the third MR
k-space dataset (STEP 208).
The method 200 can include the MRI scanner 102 acquiring the first
MR k-space dataset corresponding to the first flip angle (STEP
202), and acquiring the second MR k-space dataset of the anatomical
region of interest corresponding to a second flip angle (STEP 204).
The RF coils can emit RF pulses associated with first flip angle
and RF pulses associated with the second flip angle. The repetition
time TR can be in the order of 25 milliseconds (ms) or less. The
emitted RF pulses for each flip angle can results in corresponding
dual-echo (or multi-echo) GRE sequence. In a dual-echo GRE
sequence, the two echo times can be defined as TE1 and TE2. In
general, in a multi-echo GRE sequence, the echo times can be
defined as TE1, TE2, TE3, . . . etc. Given the application of the
frequency and phase encoding gradients, the MRI scanner 102 or the
processor 104 can use the MR signals recorded at each echo time to
fill or generate a corresponding k-space (or a corresponding
k-space dataset).
The RF coils can be used to excite an arbitrary flip angle. The MRI
scanner 102 can generate two GRE sequences with otherwise identical
structure differing only by the flip angle applied at the beginning
of the sequence. One implementation of the dual echo STAGE
acquisition would be to collect the first k-space dataset for a
first flip angle of 6 degrees and the second k-space dataset for a
second flip angle of 24 degrees. In some implementations, the MRI
scanner 102 can use other values for the first and second flip
angles. The gradients can be used to generate two or more echoes.
The MRI scanner 102 or the processor 104 can apply the inverse
Fourier transform to the MR signals (effectively the k-space for a
given echo) generate a corresponding MR image.
The method 200 can include the processor 104 generating a third MR
k-space dataset by combining the first and second MR k-space
datasets (STEP 206). The processor 104 can combine a first portion
of the first MR k-space dataset and a second portion of the second
MR k-space dataset to generate the third MR k-space dataset. The
first and second portions can be disjoint or can partially overlap
as is described in further detail below. In some implementations,
the processor 104 can combine modified versions of the first and/or
second MR k-space datasets to generate the third MR k-space
dataset.
The method 200 can include the processor 104 constructing an MR
image of the anatomical region of interest using the third MR
k-space dataset (STEP 208). The processor 104 can construct the MR
image by applying the inverse Fourier transform to the third MR
k-space dataset. The third MR k-space dataset represents a fully
sampled k-space, and as such, the reconstructed image has a higher
resolution than an MR image reconstructed either using the usual
central k-space acquired at an echo time (e.g., TE1 or TE2) of the
first flip angle or using the usual central k-space acquired at a
given echo time (e.g., TE1 or TE2) of the second flip angle.
Specifically, the method 200 allows for a higher resolution MR
image in half (or a fraction) of the usual acquisition time of the
multi-flip-angle STAGE data.
The method 200 can be implemented in various ways, for example,
depending on the number and portions of MR k-spaces acquired, the
echo times at which the MR k-spaces are acquired, preprocessing (if
any) applied to the acquired MR k-space datasets, the way portions
of the MR k-spaces are combined, or a combination thereof. Two main
approaches, namely approach I and approach II, for implementing the
method 200 are discussed in further detail with regard to FIGS.
3-8. Each of these approaches, can be implemented according to
various embodiments. Also, the MRI scanner 102 or the processor 104
can provide a user interface (UI), e.g., on the display device 108,
to allow a user to select settings for MR data acquisition and the
approach to be used to provide enhanced-resolution STAGE imaging
based on multi-flip-angle MR data. The processor 102 can cause the
MR scanner 102 to acquire MR data according to the selected
settings. For example, the selected settings can indicate the
values of the flip angles to be used, the number and/or values of
echo times at which to record MR signals, the type of output images
to be constructed, or a combination thereof. The processor 102 can
construct output MR images according to the selected STAGE imaging
approach(es), for example, among the methods described below with
regard to FIGS. 3-8.
Approach I
Referring to FIG. 3, a flowchart illustrating another method 300 of
STAGE imaging is shown, according to inventive concepts of this
disclosure. In a brief overview, the method 300 can include
acquiring, for a first flip angle, a first MR k-space dataset and a
second MR k-space dataset by scanning an anatomical region of
interest with at least two echo times TE1 and TE2 (STEP 302). The
method 300 can include acquiring, for a second flip angle, a third
MR k-space dataset and a fourth MR k-space dataset by scanning the
anatomical region of interest with the at least two echo times TE1
and TE2 (STEP 304). The method 300 can include generating a fifth
MR k-space dataset by combining the second and fourth MR k-space
datasets (STEP 306). The method 300 can include constructing an MR
image of the anatomical region of interest using the fifth MR
k-space dataset (STEP 308).
Referring to FIGS. 1 and 3, the method 300 can include the MRI
scanner 102 acquiring, for the first flip angle, the first MR
k-space dataset and the second MR k-space dataset by scanning an
anatomical region of interest with at least two echo times TE1 and
TE2 (STEP 302), and acquiring, for a second flip angle, a third MR
k-space dataset and a fourth MR k-space dataset by scanning the
anatomical region of interest with the at least two echo times TE1
and TE2 (STEP 304). The MRI scanner 102 can acquire the first MR
k-space dataset at the TE1 echo time associated with the first flip
angle, and acquire the second MR k-space dataset at the TE2 echo
time associated with the first flip angle. For instance, the MRI
scanner 102 or the corresponding RF coils can excite a dual-echo
(or a multi-echo) GRE sequence by, for example, alternating between
emitting a first MR pulse associated with the first flip angle and
a second MR pulse associated with the second flip angle. The RF
coils can receive a signal, for example, at each repetition time
TR, for first and second sets at the TE1 and TE2 echo times
associated with the first flip angle, respectively. The RF coils
can also receive a signal for third and fourth sets at the TE1 and
TE2 echo times associated with the second flip angle, respectively.
The MRI scanner 102 or the processor 104 can generate a first TE1
k-space dataset using the first set of MR signals recorded at the
TE1 echo time of the first flip angle, and a first TE2 k-space
dataset using the second set of MR signals recorded at the TE2 echo
time of the first flip angle. The MRI scanner 102 or the processor
104 can generate also a second TE1 k-space dataset using the third
set of MR signals recorded at the TE1 echo time of the second flip
angle, and a second TE2 k-space dataset using the fourth set of MR
signals recorded at the TE2 echo time of the second flip angle.
Referring to FIG. 4A, a block diagram 400A depicting an example
embodiment for acquiring and combining MR data corresponding to
multiple flip angles, according to inventive concepts of this
disclosure. The diagram 400A illustrates two blocks 402 and 404 of
a dual-echo GRE sequence having two flip angles .theta..sub.1 and
.theta..sub.2. The first block 402 corresponds to the first flip
angle .theta..sub.1, while the second block 404 corresponds to the
second flip angle .theta..sub.2. During the first block 402, the
MRI scanner 102 can acquire the first TE1 k-space dataset
representing a central portion 406 (shown in gray) of the
corresponding TE1 k-space 408, and acquire a first TE2 k-space
dataset representing a central portion 410 (shown in gray) of the
corresponding TE2 k-space 412. During the second block 404, the MRI
scanner 102 can acquire a second TE1 k-space dataset representing a
central portion 414 (shown in gray) of the corresponding TE1
k-space 416, and acquire a second TE2 k-space dataset representing
two opposite side portions 418 and 420 (shown in gray) of the
corresponding TE2 k-space 422.
In summary, the MRI scanner 102 can acquire data for the k-space
portions (or regions) 406, 410, 414, 418 and 420 shown in gray in
FIG. 4A. Acquiring these portions of the k-spaces 408, 412, 416 and
422 can make the MR data acquisition relatively faster, for
example, compared to full acquisition of the corresponding k-space.
The TE1 k-space portions 406 and 414 can have the same k-space
coverage. The TE2 k-space portion 410 can correspond to a k-space
coverage adjacent to but disjoint from (i.e., does not overlap
with) the k-space coverage of the TE2 k-space portions 418 and 420.
For instance, combining the TE2 k-space portion 410 with the TE2
k-space portions 418 and 420 can result in full k-space coverage
with higher resolution than the central k-space collected for the
first echoes. For example, let the voxels of the full k-spaces 408,
412, 416 and 422 run between -L.DELTA.k.sub.y and
(L-1).DELTA.k.sub.y along the k.sub.y axis, where .DELTA.k.sub.y
represents the step size in k-spaces along the k.sub.y axis. As
such, the total number of steps to acquire a full k-space
representing the center of k-space is equal to 2L. However,
acquiring the k-space portions 406, 410 and 414 representing the
central regions of k-spaces 408, 416 and 416, respectively, can be
achieved in 2n steps where the central regions 406, 410 and 414 are
defined by sampled k-space points with k.sub.y-coordinate k
satisfying -n.DELTA.k.sub.y.ltoreq.k.ltoreq.(n-1).DELTA.k.sub.y. In
this case, the integer n can be equal to the integer L/2. The
k-space portion 418 of the TE2 k-space 422 can be defined by the
sampled k-space points falling between -L.DELTA.k.sub.y and
-(n+1).DELTA.k.sub.y along the k.sub.y axis. The k-space portion
420 of the TE2 k-space 422 can be defined by the sampled k-space
points falling between n.DELTA.k.sub.y and (L-1).DELTA.k.sub.y
along the k.sub.y axis.
Let S.sub..theta.1,TE2 be the k-space 412 corresponding to the
first flip angle .theta..sub.1 and the echo time TE2, and let
S.sub..theta.2,TE2 be the k-space 422 corresponding to the second
flip angle .theta..sub.2 and the echo time TE2. In generating the
TE2 k-space dataset corresponding to the k-space portion 410, the
MRI scanner 102 can acquire data points S.sub..theta.1,TE2(k) for
-n.DELTA.k.sub.y.ltoreq.k.ltoreq.(n-1).DELTA.k.sub.y along the
k.sub.y axis. Also, in generating the TE2 k-space dataset
corresponding to the k-space portions 418 and 420, the MRI scanner
102 can acquire only data points S.sub..theta.2,TE2(k) for
-L.DELTA.k.sub.y.ltoreq.k.ltoreq.-(n+1).DELTA.k.sub.y and
n.DELTA.k.sub.y.ltoreq.k.ltoreq.(L-1).DELTA.k.sub.y.
Referring now to FIGS. 3 and 4A, the method 300 can include
generating a fifth MR k-space dataset by combining the second and
fourth MR k-space datasets (STEP 306). The processor 104 or the TE2
k-space combining module 424 can combine the TE2 k-space dataset
acquired at STEP 302 for the first flip angle .theta..sub.1 and the
TE2 k-space dataset acquired at STEP 304 for the second flip angle
.theta..sub.2 to generate the fifth MR k-space dataset. The TE2
k-space combining module 424 can be a component of the MR imaging
system 100 or the MRI scanner 102. The TE2 k-space combining module
424 can be a software component executable by the processor 104, a
hardware component or circuit, or a combination of software and
hardware components. Let S.sub..theta.1,.theta.2,TE2 be the final
high resolution k-space corresponding to the fifth MR k-space
dataset generated by combining the TE2 k-space dataset acquired at
STEP 302 and corresponding to the first flip angle .theta..sub.1
and the echo time TE2, and the TE2 k-space dataset acquired at STEP
304 and corresponding to the second flip angle .theta..sub.2 and
the echo time TE2. The processor 104 or the TE2 k-space combining
module 424 can generate the final full resolution k-space dataset
as:
.theta..theta..times..times..function..theta..times..times..function..tim-
es..times..times..times..DELTA..times..ltoreq..ltoreq..times..DELTA..times-
..theta..times..times..function..times..times..times..times..DELTA..times.-
.ltoreq..ltoreq..times..DELTA..times..theta..times..times..function..times-
..times..times..times..times..DELTA..times..ltoreq..ltoreq..times..DELTA..-
times. ##EQU00001##
The fifth k-space dataset as defined in equation (1) represents a
full k-space coverage for the high resolution reconstruction of
S.sub..theta.1,.theta.2,TE2 that is formed by combining the TE2
k-space dataset corresponding to the k-space portion 410 and the
TE2 k-space dataset corresponding to the k-space portions 418 and
420.
The fifth MR k-space dataset represents higher resolution MR data
compared to the first and second MR k-space datasets (i.e., the TE2
k-space datasets acquired at STEPs 302 and 304) used to generate
the MR k-space dataset because it now has a total of 2L k-space
data points. Specifically, a MR image that represents the inverse
Fourier transform of the fifth MR k-space dataset has a higher
resolution than an MR image constructed using either of the TE2
k-space datasets acquired at STEP 302 or STEP 304. The MRI scanner
102 can employ the combining of TE2 k-space datasets corresponding
to multiple flip angles to generate relatively high resolution
(e.g., compared to the resolution of the acquired MR data)
susceptibility weighted (SWI) images, high resolution true-SWI
(tSWI) images, high resolution quantitative susceptibility mapping
(QSM) images, or a combination thereof, among others.
The module 426 can be a component of the MR imaging system 100 or
the MRI scanner 102 configured to generate a T1MAP image, proton
density map (PDMAP) image or enhanced T1 weighted (T1WE) image
using the TE1 k-space dataset corresponding to the k-space portion
406 and the TE1 k-space dataset corresponding to the k-space
portion 414. Specifically, the module 426 can generate the T1MAP
image, the PDMAP image or the T1WE image as described in U.S.
patent Ser. No. 15/659,353 entitled "SYSTEMS AND METHODS FOR
STRATEGICALLY ACQUIRED GRADIENT ECHO IMAGING." The T1WE, T1Map or
PDMAP images generated by module 426 have higher signal-to-noise
ratio (SNR) compared to corresponding images generated using TE1
k-space data associated with a single flip angle. The module 426
can be a software component executable by the processor 104, a
hardware component or circuit, or a combination of software and
hardware components. The module 426 can be a component of the MRI
system 100 or the MR scanner 102.
In some implementations, the MRI scanner 102 or the imaging system
100 may acquire only TE2 k-space datasets (e.g., datasets
corresponding to k-spaces portions 410, 418 and 420) at STEPs 302
and 304, and generate the k-space S.sub..theta.1,.theta.2,TE2. For
instance, in applications where the goal is to generate SWI, tSWI
or QSM images, the imaging system 100 may omit acquiring TE1
k-space datasets at STEPs 302 and 304.
The method 300 can include the MRI scanner 102 or the processor 102
reconstructing an MR image of the anatomical region of interest
using the fifth MR k-space dataset generated at STEP 306 (STEP
308). The processor 104 can apply an inverse Fourier transform to
the generated k-space S.sub..theta.1,.theta.2,TE2 (or the k-space
dataset S.sub..theta.1,.theta.2,TE2(k)) to generate the MR image
Y.sub..theta.1,.theta.2,TE2. As discussed above, the generated MR
image has a higher resolution than an MR image constructed using
only the TE2 k-space dataset acquired at STEP 302 or an MR image
constructed using only the TE2 k-space dataset acquired at STEP
304. The constructed MR image can be processed or used to generate
an SWI image, tSWI image, or a QSM image, among others.
Referring now to FIG. 4B, a block diagram 400B depicting another
example embodiment for acquiring and combining MR data
corresponding to multiple flip angles is shown, according to
inventive concepts of this disclosure. The diagram 400B is similar
to the block diagram 400A except that in the block diagram 400B the
TE2 k-space portion 410 of the TE2 k-space 412 (corresponding to
the second flip angle .theta..sub.1) partially overlaps with the
TE2 k-space portions 418 and 420 of the TE2 k-space 422
(corresponding to the second flip angle .theta..sub.2).
Specifically, the TE2 k-space portion 410 can include the boundary
regions or segments 428 and 430 (shown in light gray). The TE2
k-space portion 418 of the TE2 k-space 422 can include the boundary
region or segment 432, which overlaps with the boundary region or
segment 428 of the TE2 k-space 412. The TE2 k-space portion 420 of
the TE2 k-space 422 can include the boundary region or segment 434,
which overlaps with the boundary region or segment 430 of the TE2
k-space 412. The boundary regions or segments 428 and 432 (shown in
light gray) can be represented by voxels having indices k between
-n.DELTA.k.sub.y and (-n+p-1).DELTA.k.sub.y along the k.sub.y axis.
Also, the boundary regions or segments 430 and 434 (shown in light
gray) can be represented by voxels having indices k between
(n-p).DELTA.k.sub.y and (n-1).DELTA.k.sub.y along the k.sub.y axis.
The boundary segments 428, 430, 432 and 434 can be viewed as having
a width equal to p.DELTA.k.sub.y, where p is an integer of
choice.
In terms of the MR data acquisition, the MRI scanner 102 can
acquire TE2 k-space data S.sub..theta.1,TE2(k) for
-n.DELTA.k.sub.y.ltoreq.k.ltoreq.(n-1).DELTA.k.sub.y along the
k.sub.y axis during the first block 402 of the dual-echo GRE
sequence. During the second block 404 of the dual-echo GRE
sequence, the MRI scanner 102 can acquire TE2 k-space data
S.sub..theta.2,TE2(k) for
-L.DELTA.k.sub.y.ltoreq.k.ltoreq.(-n+p-1).DELTA.k.sub.y and
(n-p).DELTA.k.sub.y.ltoreq.k.ltoreq.(L-1).DELTA.k.sub.y along the
k.sub.y axis. As such, the k-space coverage of the TE2 k-space data
Y.sub..theta.1,TE2(k) and the k-space coverage of the TE2 k-space
data S.sub..theta.2,TE2(k) have overlapping data corresponding to
the pair of boundary segments 428 and 432 and the pair of boundary
segments 430 and 434. The data overlap can help smooth the
transition from TE2 k-space data S.sub..theta.1,TE2(k) to TE2
k-space data S.sub..theta.2,TE2(k) when combined and used to
generate an MR image.
The MRI scanner 102 or the processor 104 can combine the datasets
S.sub..theta.1,TE2(k) and S.sub..theta.2,TE2(k) acquired at STEPs
302 and 304 as:
.theta..theta..times..times..function..theta..times..times..function..tim-
es..theta..times..times..function..times..theta..times..times..function..t-
heta..times..times..function..times..theta..times..times..function..times.-
.theta..times..times..function..theta..times..times..function..times..time-
s..times..DELTA..times..ltoreq..ltoreq..times..DELTA..times..times..times.-
.DELTA..times..ltoreq..ltoreq..times..DELTA..times..times..times..times..D-
ELTA..times..ltoreq..ltoreq..times..DELTA..times..times..times..times..DEL-
TA..times..ltoreq..ltoreq..times..DELTA..times..times..times..times..DELTA-
..times..ltoreq..ltoreq..times..DELTA..times. ##EQU00002## As
depicted in equation (2), the MRI scanner 102 or the processor 104
can use weighted sums of S.sub..theta.1,TE2(k) and
S.sub..theta.2,TE2(k) to determine S.sub..theta.1,.theta.2,TE2(k)
within each voxel of the overlap regions or segments. In equation
(2), the coefficients a.sub.1 and a.sub.2 can be defined as
.times..DELTA..times..times..DELTA..times. ##EQU00003## and
a.sub.2=1-a.sub.1. Also, the coefficients b.sub.1 and b.sub.2 can
be defined as b.sub.1=1-b.sub.2, and
.times..DELTA..times..times..DELTA..times. ##EQU00004## The
weighted sum approach used to combine S.sub..theta.1,TE2(k) and
S.sub..theta.2,TE2(k) within the overlap regions or segments allows
for a smooth transition between S.sub..theta.1,TE2(k) data and
S.sub..theta.2,TE2(k) data. In some implementations, other
mathematical weightings for the coefficients could be used.
In some implementations, the MRI scanner 102 or the processor 104
can use phase difference between the acquired dataset (e.g.,
corresponding to region 406 of FIG. 4A or regions 406, 436 and 438
of FIG. 4B) for the first TE1 k-space S.sub..theta.1,TE2 408 and
the acquired dataset (e.g., corresponding to region 410 of FIG. 4A
or regions 410, 428 and 430 of FIG. 4B) for the first TE2 k-space
S.sub..theta.1,TE2 412 as well as the acquired dataset (e.g.,
corresponding to region 414 of FIG. 4A or regions 414, 440 and 442
of 4B) for the second TE1 k-space S.sub..theta.2,TE1 416 to
determine the central portion 444 of the second TE2 k-space
S.sub..theta.2,TE2 422. Note that the phase difference between the
TE1 k-space and the TE2 k-space is the same for both flip angles
.theta..sub.1 and 02. That is the phase difference between the
first TE1 k-space S.sub..theta.1,TE1 408 and the first TE2 k-space
S.sub..theta.1,TE2 412 is the same as the phase difference between
the second TE1 k-space S.sub..theta.2,TE1 416 and the second TE2
k-space S.sub..theta.2,TE2 422.
The MRI scanner 102 or the processor 104 can apply the inverse
Fourier transform to the acquired dataset for the first TE1 k-space
S.sub..theta.1,TE1 408 to generate a corresponding MR image. The
MRI scanner 102 or the processor 104 can apply the inverse Fourier
transform to the acquired dataset for the first TE2 k-space
S.sub..theta.1,TE2 412 to generate a corresponding MR image. The
MRI scanner 102 or the processor 104 can complex divide the MR
image corresponding to the acquired dataset for the first TE1
k-space S.sub..theta.1,TE2 408 by the MR image corresponding to the
acquired dataset for the first TE2 k-space S.sub..theta.1,TE2 412
to determine a phase difference and a T2* weighting factor. The MRI
scanner 102 or the processor 104 can apply inverse Fourier
transform to the acquired dataset (e.g., region 414 in FIG. 4A or
regions 414, 440 and 442 of FIG. 4B) for the second TE1 k-space
S.sub..theta.2,TE1 416 to generate a corresponding MR image, and
then adjust the MR image corresponding to the acquired dataset for
the second TE1 k-space S.sub..theta.2,TE1 416 using the phase
difference and the T2* weighting factor. Specifically, the MRI
scanner 102 or the processor 104 can multiply the MR image
corresponding to the acquired dataset for the second TE1 k-space
S.sub..theta.2,TE1 416 by the T2* weighting factor and the phase
term e.sup.-i.PHI.(y), where .PHI.(y) represents the phase
difference between the MR image corresponding to the acquired
dataset for the first TE1 k-space S.sub..theta.1,TE1 408 and the MR
image corresponding to the acquired dataset for the first TE2
k-space S.sub..theta.1,TE2 412.
The MRI scanner 102 or the processor 104 can apply the Fourier
transform to the adjusted MR image to generate corresponding
k-space data. The k-space data corresponding to the adjusted MR
image, which can be viewed as a central extrapolation of the
acquired MR k-space dataset for the second TE1 k-space
S.sub..theta.2,TE1 416, represents an estimate of the central
region 428 or 444 of the second TE2 k-space S.sub..theta.2,TE2 422.
The MRI scanner 102 or the processor 104 can combine the k-space
data corresponding to the adjusted MR image and the acquired
k-space dataset (e.g., regions 418 and 420 of FIG. 4A or regions
418, 420, 432 and 434 of FIG. 4B) for the second TE2 k-space
S.sub..theta.2,TE2 422. The combining of these k-space datasets
results in full data for the second TE2 k-space S.sub..theta.2,TE2
422.
The k-space dataset corresponding to the adjusted MR image and the
acquired k-space dataset (e.g., regions 418 and 420 of FIG. 4A or
regions 418, 420, 432 and 434 of FIG. 4B) for the second TE2
k-space S.sub..theta.2,TE2 422 may be disjoint (e.g., no overlap).
In such cases, the MRI scanner 102 or the processor 104 can combine
the two k-space datasets as described in relation to equation (1),
except that the k-space dataset corresponding to the adjusted MR
image is used instead of S.sub..theta..sub.1.sub.,TE2(k). However,
if the k-space dataset corresponding to the adjusted MR image and
the acquired k-space dataset (e.g., regions 418 and 420 of FIG. 4A
or regions 418, 420, 432 and 434 of FIG. 4B) for the second TE2
k-space S.sub..theta.2,TE2 422 overlap, the MRI scanner 102 or the
processor 104 can combine the two k-space datasets as described in
relation to equation (2), except that the k-space dataset
corresponding to the adjusted MR image is used instead of
S.sub..theta..sub.1.sub.,TE2(k).
Approach II
Referring now to FIG. 5, a block diagram 500 depicting another
example embodiment for acquiring and combining MR data
corresponding to multiple flip angles is shown, according to
inventive concepts of this disclosure. The diagram 500 illustrates
two blocks 502 and 504 of a dual-echo GRE sequence. The first block
502 corresponds to the first flip angle .theta..sub.1, while the
second block 504 corresponds to the second flip angle
.theta..sub.2. During the first block 504 of the dual-echo GRE
sequence, the MRI scanner 102 can acquire a first TE1 k-space
dataset representing a central portion 506 of the corresponding TE1
k-space 508, and acquire a first TE2 k-space dataset representing a
first side portion 510 of the corresponding TE2 k-space 512. The
MRI scanner 102 can omit (or skip) acquiring TE2 k-space data
corresponding to the k-space region 514 (shown in white). During
the second block 504, the MRI scanner 102 can acquire a second TE1
k-space dataset representing a central portion 516 of the
corresponding TE1 k-space 518, and acquire a second TE2 k-space
dataset representing a second side portion 520 of the corresponding
TE2 k-space 522. The MRI scanner 102 can omit (or skip) acquiring
TE2 k-space data corresponding to the k-space region 526 (shown in
white). The first and second side portions 510 and 520 can be
opposite to one another.
In some implementations, the first and second side portions 510 and
520 can partially overlap. For instance, the first side portion 510
can include a first boundary or overlap region 524 (shown in light
gray), and the second side portion 520 can include a second
boundary or overlap region 528 (shown in light gray). The first and
second boundary or overlap regions 524 and 528 can fully overlap
with one another. The first side portion 510 can be defined as the
set of points (k) where
-L.DELTA.k.sub.y.ltoreq.k<q.DELTA.k.sub.y, and the first
boundary or overlap region 524 can be defined as the set of points
(k) where -q.DELTA.k.sub.y.ltoreq.k<q.DELTA.k.sub.y. Here q is
an integer and the width of the first boundary or overlap region
524 is equal to 2q.DELTA.k.sub.y. The second side portion 520 can
be defined as the set of points (k) where
-q.DELTA.k.sub.y.ltoreq.k.ltoreq.(L-1).DELTA.k.sub.y, and the
second boundary or overlap region 528 can be defined as (similar to
the first overlap region 524) the set of points (k) where
-q.DELTA.k.sub.y.ltoreq.k<q.DELTA.k.sub.y.
The TE2 k-space combining module 530 can be a component of the
imaging system 100 or the MRI scanner 102. The TE2 k-space
combining module 530 can be a software component (e.g., executable
by the processor 104), a hardware component or circuit, or a
combination of both. The module 532 can be similar to the module
426 of FIGS. 4A and 4B. Similar to FIGS. 4A and 4B, let
S.sub..theta.1,TE2 be the TE2 k-space 512 corresponding to the
first flip angle .theta..sub.1, and let S.sub..theta.2,TE2
represent the TE2 k-space 522 corresponding to the second flip
angle .theta..sub.2. The TE2 k-space dataset corresponding to the
k-space portion 510 can be defined as the S.sub..theta.1,TE2(k) for
-L.DELTA.k.sub.y.ltoreq.k<q.DELTA.k.sub.y. The TE2 k-space
dataset corresponding to the k-space portion 520 can be defined as
the S.sub..theta.2,TE2(k) for
-q.DELTA.k.sub.y.ltoreq.k<(L-1).DELTA.k.sub.y. The TE2 k-space
combining module 530 (or the processor 104) can combine these
k-space datasets (STEP 306 of method 300) to generate the fifth
k-space dataset S.sub..theta.2,.theta.2,TE2(k) as:
.theta..theta..times..times..function..theta..times..times..function..tim-
es..theta..times..times..function..times..theta..times..times..function..t-
heta..times..times..function..times..times..times..DELTA..times..ltoreq.&l-
t;.times..DELTA..times..times..times..DELTA..times..ltoreq.<.times..DEL-
TA..times..times..times..times..DELTA..times..ltoreq.<.times..DELTA..ti-
mes. ##EQU00005## In equation (3), the coefficients c.sub.1 and
c.sub.2 can be set equal to 0.5.
The data overlap along the pair of boundary segments 524 and 528
can result in phase discrepancy within the boundary segments when
combining the TE2 k-space data S.sub..theta.1,TE2(k) and the TE2
k-space data S.sub..theta.2,TE2(k) and using
S.sub..theta.1,.theta.2,TE2(k) to generate an MR image. To address
this issue, the MRI scanner 102 or the processor 104 can adjust the
phase data for any of the MR images corresponding to
S.sub..theta.1,TE2(k) and S.sub..theta.2,TE2(k) prior to combining
the k-space datasets S.sub..theta.1,TE2(k) and
S.sub..theta.2,TE2(k) within the overlap regions or segments.
Specifically, prior to combining the datasets S.sub..theta.1,TE2(k)
and S.sub..theta.2,TE2(k) within the overlap regions or segments,
the MRI scanner 102 or the processor 104 can compare the phase of
the images that result from each of these k-space datasets. The MRI
scanner 102 or the processor 104 can fill the regions outside the
collected data for each k-space with zeroes until the k-space is
full and then take the inverse Fourier transform of
S.sub..theta.1,TE2(k) and S.sub..theta.2,TE2(k) to generate two
images U.sub..theta.1,TE2(y) and U.sub..theta.2,TE2(y). The MRI
scanner 102 or the processor 104 can compare the phase information
of the images U.sub..theta.1,TE2(y) and U.sub..theta.2,TE2(y), for
example, by complex dividing U.sub..theta.2,TE2(y) by
U.sub..theta.1,TE2(y) to determine e.sup.i.phi.(y) where .phi.(y)
represents the phase difference. The MRI scanner 102 or the
processor 104 can adjust the phase information of
U.sub..theta.2,TE2(y) by computing
V.sub..theta.2,TE2(y)=e.sup.-i.phi.(y)U.sub..theta.2,TE2(y) so that
both images U.sub..theta.1,TE2(y) and V.sub..theta.2,TE2(y) have
the same phase information. The MRI scanner 102 or the processor
104 can apply the Fourier transform to V.sub..theta.2,TE2(y) to
compute the corresponding k-space T.sub..theta.2,TE2(k). The
k-space T.sub..theta.2,TE2(k) can be viewed as a modified (or
processed) version of S.sub..theta.2,TE2(k).
Now, the MRI scanner 102 or the processor 104 can generate the
k-space S.sub..theta.1,.theta.2,TE2(k) by combining the k-space
datasets S.sub..theta.1,TE2(k), S.sub..theta.2,TE2(k) and
T.sub..theta.2,TE2(k) as:
.theta..theta..times..times..function..theta..times..times..function..tim-
es..theta..times..times..function..times..theta..times..times..function..t-
heta..times..times..function..times..times..times..DELTA..times..ltoreq.&l-
t;.times..DELTA..times..times..times..DELTA..times..ltoreq.<.times..DEL-
TA..times..times..times..times..DELTA..times..ltoreq.<.times..DELTA..ti-
mes. ##EQU00006## Equation (4) is similar to equation (3), except
for the use of T.sub..theta.2,TE2(k) instead of
S.sub..theta.2,TE2(k) within the overlap boundary region where
-q.DELTA.k.sub.y.ltoreq.k<q.DELTA.k.sub.y. Specifically, in
equation (4), the coefficients d.sub.1 and d.sub.2 can be set equal
to 0.5 similar to the coefficients c.sub.1 and c.sub.2 of equation
(3). Correcting for any phase difference eliminates or mitigates
undesired image artifacts due to such phase difference. In some
implementations, the MRI scanner 102 or the processor 104 can
adjust the phase information of U.sub..theta.1,TE2(y) (instead of
U.sub..theta.1,TE2(y)) and use the corresponding k-space
T.sub..theta.1,TE2(k) instead of T.sub..theta.2,TE2(k) in equation
(4). The processor 104 can apply an inverse Fourier transform to
the generated k-space data S.sub..theta.1,.theta.2,TE2(k) to
generate the MR image Y.sub..theta.1,.theta.2,TE2. Approach III
Referring to FIG. 6, a flowchart illustrating another method 600 of
STAGE imaging using MR data corresponding to multiple flip angles
is shown, according to inventive concepts of this disclosure. The
method 600 can include acquiring, for a first flip angle, a first
MR k-space dataset of an anatomical region of interest using a
first echo of a predefined type (STEP 602), and acquiring, for a
second flip angle, a second MR k-space dataset of the anatomical
region of interest using a second echo of a predefined type (STEP
604). The method 600 can include computing a first MR image
representing an inverse Fourier transform of the first MR k-space
dataset and a second MR image representing an inverse Fourier
transform of the second MR k-space dataset (STEP 606). The method
600 can include computing a third MR image representing a linear
transformation of the first MR image and a fourth MR image
representing a linear transformation of the second MR image (STEP
608). The method 600 can include generating a third MR k-space
dataset using the third MR image and a fourth MR k-space dataset
using the fourth MR image (STEP 610). The method 600 can include
combining the third and fourth MR k-space datasets to generate a
fifth MR k-space dataset (STEP 612), and reconstructing a fifth MR
image of the anatomical region of interest using the fifth k-space
dataset (STEP 614).
The method 600 can include the MRI scanner acquiring, for a first
flip angle, a first MR k-space dataset S.sub..theta.1,TEn(k) of an
anatomical region of interest using a first echo of a predefined
type such as TEn (STEP 602), and acquiring, for a second flip
angle, a second MR k-space dataset S.sub..theta.2,TEn(k) of the
anatomical region of interest using a second echo of the predefined
type (STEP 604). As discussed above with regard to FIGS. 3-5, the
MRI scanner 102 can trigger a first dual-echo (or multi-echo) GRE
sequence having a first block (e.g., block 402 or 502) associated
with the first flip angle .theta..sub.1 and a second block (e.g.,
block 404 or 504) associated with the second flip angle
.theta..sub.2. The MRI scanner 102 can acquire the first MR k-space
dataset S.sub..theta.1,TEn(k) during the first block of the dual
echo (or multi-echo) GRE sequence, and acquire the second MR
k-space dataset S.sub..theta.2,TEn(k) during the second block of
the dual echo (or multi-echo) GRE sequence. Each of the first and
second MR k-space datasets S.sub..theta.1,TEn(k) and
S.sub..theta.2,TEn(k) can represent a portion of the corresponding
k-space (e.g., does not fully cover the corresponding k-space but
only a portion thereof) as discussed with regard to FIGS. 3-5.
Unlike method 300 where the first and second MR k-space datasets
are TE2 k-space datasets, here the first and second MR k-space
datasets can be both TE1 k-space datasets, both TE2 k-space
datasets, both TE3 k-space datasets, both TE4 k-space datasets, or
a combination of k-space datasets associated with different types
of echo times, among others. MR images corresponding to k-space
datasets associated with different flip angles can have distinct
visual characteristics. Specifically, the intensities (or average
intensities) associated with different tissue types and/or the
contrast between the different tissue types may vary in MR images
corresponding to distinct flip angles.
Referring to FIG. 7, various MR brain images corresponding to
distinct flip angles for the TE1 data are shown. The brain image
702 is generated using TE1 data corresponding to a flip angle equal
to 24 degrees, the brain image 704 is generated using TE1 data
corresponding to a flip angle equal to 6 degrees, and the brain
image 706 is generated using TE1 data corresponding to a flip angle
equal to 2 degrees. In the brain image 702 corresponding to the
high flip angle 24 degrees, the white matter has the highest
intensity, the gray matter has the next highest intensity, and the
cerebral spinal fluid (CSF) has the lowest intensity. When
considering the brain image 706 or 708 corresponding to the
2.degree. flip angle, the CSF has the highest intensity, the gray
matter has the next highest intensity, and the white matter has the
lowest intensity. This is not true, however, for the brain image
704 corresponding to the 6.degree. flip angle. The difference in
signal intensities between the MR image 702 and the MR image 706 or
708 is due to the fact that the water content dominates the low
flip angle image 706 or 708 and the CSF is basically 100% water,
the gray matter is about 84% water, and the white matter is about
68% water.
Unlike brain images 706 and 708, the brain image 704 corresponding
to the 6.degree. flip angle does not show contrast in intensities
between the three brain regions opposite to corresponding contrast
shown in brain image 702. However, the MRI scanner 102 or the
processor 104 can use acquired datasets corresponding to flip
angles 24.degree. and 6.degree. to generate T1maps and PDmaps as
described in U.S. patent Ser. No. 15/659,353 entitled "SYSTEMS AND
METHODS FOR STRATEGICALLY ACQUIRED GRADIENT ECHO IMAGING." Once the
T1maps and PDmaps are generated, the MRI scanner 102 or the
processor 104 can simulate or generate the synthetic image for any
flip angle. Hence, the MRI scanner 102 or the processor 104 can
generate the synthetic image for the 2.degree. flip angle, such as
image 708, using the T1maps and PDmaps. The MRI scanner 102 or the
processor 104 can use the synthetic image for the 2.degree. flip
angle, instead of the image corresponding to the 6.degree. flip
angle, in the rest of the steps of the method 600. As illustrated
in FIG. 7, the simulated or synthetic image 708 corresponding to a
2.degree. flip angle looks identical to the image 706 actually
acquired at a 2.degree. flip angle except that it may have higher
signal-to-noise ratio.
The variation, based on the flip angle, in contrast and signal
intensities for the various brain regions calls for processing MR
datasets corresponding to distinct flip angles before combining
such datasets. For instance, by subtracting from each of the MR
images 702 and 708 (or 706, e.g., if the small flip angle is equal
to 2.degree.) the corresponding baseline, the resulting images
would have opposite contrasts. As such, scaling one of the images
(with removed baseline) can cause the two images to look similar.
Specifically, by applying proper negative scaling to one of the MR
images (after baseline subtraction), the contrast between any two
tissue types can be made similar (e.g., to some extent) across the
two MR images. With respect to combining MR k-space datasets
corresponding to distinct flip angles, the second echo MR k-space
datasets can be modified, before combining them, such that the
corresponding MR images have relatively similar contrasts between
different tissue types. Such processing prior to combining the MR
k-space datasets can lead to a reduction of artifacts in the MR
image obtained from the combined k-space.
Referring back to FIG. 6, the method 600 can include the MRI
scanner 102 or the processor 104 computing a first MR image
representing the inverse Fourier transform of the first MR k-space
dataset, and a second MR image representing the inverse Fourier
transform of the second MR k-space dataset (STEP 606). Transforming
the first and second MR k-space datasets to image data can allow
for determining the processing to be performed on the MR images so
that the processed images have similar visual characteristics for
the different tissue types. For instance, the transformation from
the k-space domain to the image domain (or the spatial domain)
allows for determining the adjustment to be made to the
intensities, the contrasts or other visual characteristics of the
MR images so that the various tissue types look similar in the
adjusted MR images.
The method 600 can include the MRI scanner 102 or the processor 104
computing a third image representing a linear transformation of the
first MR image (or a linear transformation of another image
associated with the first MR image), and fourth MR image
representing a linear transformation of the second MR image (STEP
608). For example, let X.sub..theta.1,TEn(y) be the MR image
representing the inverse Fourier Transform of S.sub..theta.1,TEn(k)
and let X.sub..theta.2,TEn(y) be the MR image representing the
inverse Fourier transform of S.sub..theta.2,TEn(k). As discussed
above with regard to FIG. 7, if the first flip angle .theta..sub.1
is not small enough (e.g., 6.degree.), the MRI scanner 102 or the
processor 104 can use a synthetic (or simulated) MR image
X.sub..theta.1',TEn(y) for a smaller flip angle .theta..sub.1'
(e.g., 2.degree.) instead of the MR image X.sub..theta.1,TEn(y). In
the following, the angle .theta..sub.s represents either the flip
angle .theta..sub.1 or the flip angle .theta..sub.1' and
X.sub..theta.s,TEn(y) represents either the MR image
X.sub..theta.1,TEn(y) or the MR image X.sub..theta.1',TEn(y)
depending on, for example, how small is the flip angle
.theta..sub.1 and which MR image is used for further processing.
The MRI scanner 102 or the processor 104 can determine for each of
the MR images X.sub..theta.s,TEn(y) and X.sub..theta.2,TEn(y) a
respective baseline value, such as a value for some particular
tissue of interest. Let .beta..sub.1 be the baseline value for
X.sub..theta.s,TEn(y) and let (32 be the baseline value for
X.sub..theta.2,TEn(y). The MRI scanner 102 or the processor 104 can
compute .beta..sub.1 as the CSF value in X.sub..theta.s,TEn(y), and
can compute (32 as the CSF value in X.sub..theta.2,TEn(y). The MRI
scanner 102 or the processor 104 can subtract, from each of the MR
images X.sub..theta.s,TEn(y) and X.sub..theta.2,TEn(y) the
respective baseline value to compute
Z.sub..theta.s,TEn(y)=|X.sub..theta.s,TEn(y)-.beta..sub.1| and
Z.sub..theta.2,TEn(y)=X.sub..theta.2,TEn(y)-.beta..sub.2.
The MRI scanner 102 or the processor 104 can determine for at least
one of the MR images (e.g., after baseline subtraction) a
respective scaling factor .alpha.. The scaling factor .alpha. can
be viewed as a proportionality value between the peak intensity of
one MR image and the peak intensity in the other MR image. In some
implementations, the MRI scanner 102 or the processor 104 can
determine the scaling factor as
.alpha..function..theta..function..theta. ##EQU00007## The MRI
scanner 102 or the processor 104 can transform the MR image
X.sub..theta.s,TEn(y) to
Z.sub..theta.s,TEn(y)=X.sub..theta.s,TEn(y)-.beta..sub.1 and
transform the MR image X.sub..theta.2,TEn(y) to
Z'.sub..theta.2,TEn(y)=.alpha.Z.sub..theta.2,TEn(y)=a
(X.sub..theta.2,TEn(y)-.beta..sub.2).
In general the MRI scanner 102 or the processor 104 can transform
the MR image X.sub..theta.s,TEn to Z.sub..theta.s,TEn=.alpha..sub.1
(X.sub..theta.s,TEn-.beta..sub.1) and transform the MR image
X.sub..theta.2,TEn to Z'.sub..theta.2,TEn=.alpha..sub.2
(X.sub..theta.2,TEn-.beta..sub.2), where .alpha..sub.1 and
.alpha..sub.2 represent two scaling factors. The MRI scanner 102 or
the processor 104 can determine the parameters .alpha..sub.1,
.alpha..sub.2, .beta..sub.1 and .beta..sub.2 such that the adjusted
(or processed) MR images Z.sub..theta.s,TEn and Z'.sub..theta.2,TEn
have similar visual characteristics (e.g., similar intensities for
each type of tissue). For example, the MRI scanner 102 or the
processor 104 can determine the parameters .alpha..sub.1,
.alpha..sub.2, .beta..sub.1 and .beta..sub.2 such that the MR
images Z.sub..theta.s,TEn and Z'.sub..theta.s,TEn have equal
maximum intensities, equal minimum intensity, equal maximum
contrast, or equal maximum contrast between a given pair of tissue
types, among others.
The method 600 can include the MRI scanner 102 or the processor 104
generating a third k-space dataset corresponding to the third MR
image representing the linear transformation of the first MR image
(or the linear transformation of another MR image associated with
the first MR image), and a fourth k-space dataset corresponding to
the fourth MR image representing the linear transformation of the
second MR image (STEP 610). For instance, the MRI scanner 102 or
the processor 104 can generate the third MR k-space dataset
W.sub..theta.s,TEn(k) as the Fourier transform of the MR image
Z.sub..theta.s,TEn(y), and can generate the fourth MR k-space
dataset W.sub..theta.2,TEn(k) as the Fourier transform of the MR
image Z'.sub..theta.2,TEn(y). In general, the MRI scanner 102 or
the processor 104 can transform the third and fourth MR images back
to the k-space domain.
The method 600 can include the MRI scanner 102 or the processor 104
combining the third and fourth MR k-space datasets to generate a
fifth MR k-space dataset (STEP 612). The MRI scanner 102 or the
processor 104 can combine the third and fourth MR k-space datasets
in a similar way as discussed above with regard to FIGS. 2-5. For
example, the third and fourth MR k-space datasets can correspond to
distinct k-space portions as those described in, or discussed with
regard to FIGS. 4A, 4B and 5. Also, as discussed above with regard
to FIGS. 2-5, the k-space portions corresponding to the third and
fourth MR k-space datasets can partially overlap (e.g., similar to
partial overlap described in, and discussed with regard to, FIGS.
4B and 5). The combining of the third and fourth MR k-space
datasets W.sub..theta.s,TEn(k) and W.sub..theta.2,TEn(k) can be
performed according to any of the techniques described above in
equations (1)-(4) except for the fact that the third, fourth and
fifth k-space datasets can be associated with any echo time and are
not restricted to TE2. For instance, where W.sub..theta.s,TEn(k)
and W.sub..theta.2,TEn(k) are combined similarly to equation (3)
above, the MRI scanner 102 or the processor 104 can generate a new
fully covered k-space W.sub..theta.s, .theta.2,TEn(k) as:
.theta..theta..function..theta..function..times..theta..function..times..-
theta..function..theta..function..times..times..times..DELTA..times..ltore-
q.<.times..DELTA..times..times..times..DELTA..times..ltoreq.<.times.-
.DELTA..times..times..times..times..DELTA..times..ltoreq.<.times..DELTA-
..times. ##EQU00008## where the constants e.sub.1 and e.sub.2 are
set to 0.5 similar to the constants c.sub.1 and c.sub.2 of equation
(3). Equation (3) of Approach II can be viewed as a special case of
equation (5) with n=2 and .theta..sub.s equal to .theta..sub.1.
In some implementations, the MRI scanner 102 or the processor 104
can apply phase adjustment before combining W.sub..theta.1,TE1(k)
and W.sub..theta.2,TE1(k) within the overlap region as discussed
above with regard to equation (4). For instance, the MRI scanner
102 or the processor 104 can complex divide Z.sub..theta.s,TEn(y)
by Z'.sub..theta.2,TEn(y) to determine e.sup.i.PHI.(y) where
.PHI.(y) represents the phase difference. The MRI scanner 102 or
the processor 104 can adjust the phase information of
Z'.sub..theta.2,TEn(y) by computing
V.sub..theta.2,TEn(y)=e.sup.-.phi.(y)Z'.sub..theta.2,TEn(y) so that
both images Z.sub..theta.s,TEn(y) and V.sub..theta.2,TEn(y) have
the same phase information. The MRI scanner 102 or the processor
104 can apply the Fourier transform to V.sub..theta.2,TEn(y) to
compute the corresponding k-space T.sub..theta.2,TEn(k). The MRI
scanner 102 or the processor 104 can generate the k-space
W.sub..theta.s,.theta.2,TEs(k) by combining the k-space datasets
W.sub..theta.s,TEn(k), W.sub..theta.2,TEn(k) and
Z'.sub..theta.2,TEn(k) as:
.theta..theta..function..theta..function..times..theta..function..times..-
theta..function..theta..function..times..times..times..DELTA..times..ltore-
q.<.times..DELTA..times..times..times..DELTA..times..ltoreq.<.times.-
.DELTA..times..times..times..times..DELTA..times..ltoreq.<.times..DELTA-
..times. ##EQU00009## where the constants f.sub.1 and f.sub.2 are
set to 0.5 similar to the constants e.sub.1 and e.sub.2 of equation
(5). Equation (4) of Approach II can be viewed as a special case of
equation (6) with n=2 and .theta..sub.s equal to .theta..sub.1.
The method 600 can include the MRI scanner 102 or the processor 104
reconstructing an MR image of the anatomical region of interest
using the fifth MR k-space dataset (STEP 614). The MRI scanner 102
or the processor 104 can apply the inverse Fourier transform to the
W.sub..theta.s,.theta.2,TEn(k) dataset to reconstruct the now high
resolution complex MR image Y.sub..theta.s,.theta.2,TEn(y) of the
anatomical region of interest. One can then add back a final
constant to the image Y.sub..theta.s,.theta.2,TEn(y) equal to the
baseline value .beta..sub.1 that was originally subtracted from
X.sub..theta.s,TEn(y). Using the phase information from this image,
the MRI scanner 102 or the processor 104 can create a new HR SWI
STAGE image.
In general, the MRI scanner 102 or the processor 104 can employ
APPROACH III to generate a spin density weighted image or a T1
weighted image, when the first and second echo times associated
with the first and second k-space datasets are TE1 echo times, or
to generate a susceptibility weighted image or a quantitative
susceptibility mapping (QSM) image when the first and second echo
times associated with the first and second k-space datasets are TE2
echo times. In the case that there are multiple echoes, this
process of merging k-space data sets can be done for any or all
desired echoes. The exact implementation will depend on how k-space
is collected at each echo. An example implementation can include
breaking up each echo into an equal number of k-space lines chosen
to fill in the missing k-space lines desired for the final high
resolution image.
Referring to FIG. 8, a block diagram 800 depicting another example
approach for acquiring and combining multi-echo MR data
corresponding to multiple flip angles is shown, according to
inventive concepts of this disclosure. The block diagram 800
illustrates an example multi-echo implementation of STAGE imaging.
The MRI scanner 102 can trigger a multi-echo GRE sequence having
two blocks 802 and 804. The first multi-echo GRE sequence block 802
corresponds to a first flip angle .theta..sub.1, while the second
multi-echo GRE sequence block 804 corresponds to a second flip
angle .theta..sub.2. The second flip angle .theta..sub.2 can be
greater than the first flip angle .theta..sub.1. For example, the
first flip angle .theta..sub.1 can be equal to 6 degrees, while the
second flip angle .theta..sub.2 can be equal to 24 degrees. For
each of the multi-echo GRE sequence blocks 802 and 804, the MRI
scanner 102 can acquire partial MR k-space datasets (shown in gray
strips) at each echo time of a plurality of echo times TE1, TE2, .
. . , TEn, where n is an integer.
The implementation described in FIG. 8 can be viewed as extending
the idea of STAGE imaging to a multi-echo mode. The MRI scanner 102
or the processor 104 can use the TE1 k-space datasets across the
multi-echo GRE sequence blocks 802 and 804 to form T1 weighted
enhanced (T1WE), T1, proton spin density (PD), B.sub.1.sup.+ and/or
B.sub.1.sup.- mapping images of relatively high signal to noise
ratio (SNR). On the other hand, the MRI scanner 102 or the
processor 104 can combine the k-space datasets associated with
higher echo times (e.g., TE2, TE3, TE4, TEn), that are acquired
across the multi-echo GRE sequence blocks 802 and 804, to form
higher resolution phase images for use in creating SWI, tSWI and/or
QSM images.
The second k-space combining module 808 can combine the k-space
datasets acquired at the echo times TE2, TE3, TE4, . . . , TEn in
the multi-echo GRE sequence blocks 802 and 804 to form a higher
resolution final k-space dataset for use to generate an MR image.
For instance, the MRI scanner 102 or the processor 104 can setup
the k-space center at a certain echo (e.g., TEn in FIG. 8) other
than the first echo to dominate the susceptibility contrast. The
second k-space combining module 808 can use, for example, all
echoes other than the first echo for filling a high-resolution
k-space by a center-out phase encoding design to get an effective
phase encoding equal to nk.sub.c. The first echo and the echo for
the SWI k-space center can be encoded for the strip consisting of
the central k.sub.c+2p lines of the k-space, where p represents the
number of overlapping lines in k-space for each of the remaining
n-2 echoes. The remaining n-2 echoes of each flip angle can each be
encoded for a respective strip having k.sub.c/2+p lines of k-space.
As depicted in FIG. 8, in the first multi-echo GRE sequence block
802 corresponding to the first flip angle, the strips corresponding
to the n-2 echoes TE2 . . . TE(n-1) echoes can be arranged on one
side (e.g., to the left) of the strip corresponding to the TEn
echo. In the second multi-echo GRE sequence block 804 corresponding
to the second flip angle, the strips corresponding to the n-2
echoes TE2 . . . TE(n-1) echoes can be arranged on the other side
(e.g., to the right) of the strip corresponding to the TEn echo.
The second k-space combining module 808 can combine k-space
datasets acquired across the multi-echo GRE sequence blocks 802 and
804 according to any of the methods described above with regard to
FIGS. 2-7.
While in FIG. 8, the second k-space combining module 808 is
designed or configured to combine the k-space datasets acquired at
the echo times TE2, TE3, TE4, . . . , TEn in the multi-echo GRE
sequence blocks 802 and 804, according to a more general
implementation, the second k-space combining module 808 can use any
combination of the k-space datasets acquired at all the echo times
(including TE1) across the multi-echo GRE sequence blocks 802 and
804. The second k-space combining module 808 or the processor 104
can apply some processing (e.g., as discussed with regard to FIG.
6) to the k-space datasets or the corresponding MR images before
combining them. The MRI scanner 102 or the processor 104 can use
any of the methods or techniques discussed above with regard to
FIGS. 2-7 to generate the high resolution k-space dataset. The MRI
scanner 102 or the processor can apply inverse Fourier transform to
the high resolution k-space dataset to construct a higher
resolution (or higher SNR) MR image of the anatomical region of
interest.
In some implementations, the MRI scanner 102 or the processor 104
can execute a combination of the STAGE imaging methods or
approaches described above with regard to FIGS. 2-8. For example,
the MRI scanner 102 or the processor 104 can employ the STAGE
imaging approach described with regard to FIGS. 3-5 to enhance or
increase the resolution of echo data associated with echo times
other than TE1, and use the STAGE imaging approach to enhance or
increase the resolution of TE1 echo (or any other echo) data. The
imaging system 100 or the MRI scanner 102 can execute any of these
STAGE imaging approaches with either no or minimal increase in
execution time. However, by combining these methods using a
conventional segmented k-space acquisition as shown in FIG. 8, the
data can either be acquired in half the time, or the resolution can
be doubled (or increased) yet again, or the slice thickness cut in
half (while doubling the number of slices) at no further expense in
time. In some implementations, the imaging system 100 or the MRI
scanner 102 can implement these STAGE imaging concepts by using
short echo time separations with more echoes. In some
implementations, the imaging system 100 or the MRI scanner 102 can
apply these STAGE imaging concepts with other fast MR imaging
methods, such as parallel imaging and compressed sensing.
To validate the STAGE imaging approaches described above, one can
compare MR images constructed using these imaging approaches to
images constructed using acquired high resolution k-space data. The
comparison can allow for visualization and quantification of the
reproducibility of high resolution data by merging or combining
relatively low resolution k-space datasets corresponding to
different flip angles (see FIGS. 9 and 10).
Referring to FIG. 9, STAGE imaging simulation results for various
types of MR images are shown. The first row of images corresponds
to fully sampled original data. Specifically, the first row shows
T1 map, PD map, T1WE, simulated double inversion recovery (DIR) GM,
tSWI, and QSM images, where each of which is constructed using a
respective fully acquired k-space. The second row shows
corresponding MR images reconstructed using k-space datasets
acquired at two flip angles (6 and 24 degrees). The MR data is
acquired by scanning the brain of a patient with Sturge-Weber
syndrome (SWS) (10y1m, male). T1 map, PD map and T1WE images of the
second row have an overall SNR increase of 62.9% than the
corresponding MR images in the first row. The SNR is computed by
manually drawn multiple regions on WM region. The tSWI and QSM
images in the second row are visually very close to the
corresponding images in the first row but were acquired in an
equivalent of half the original time. The tSWI and QSM images show
minimum intensity projection (mIP) for tSWI and a maximum intensity
projection (MIP) for QSM both over 8 slices.
FIG. 10 shows STAGE imaging simulation results for tSWI images and
QSM images. FIG. 10 shows three sets of SWI and QSM data consisting
of fully-sampled MR data acting as ground truth (GT), k-space
sharing (Y.sub..theta.1,.theta.2,TE2(y)) data referring to MR
images constructed by combining various k-space datasets, and
central undersampled (U.sub..theta.1,.theta.2,TE2(y)) data. In
comparing these MR data sets, one can use a voxel based normalized
root mean square error (NRMSE),
NRMSE.sub.U,Y=sqrt((GT-U,Y.sub..theta.1,.theta.2,TE2(y)).sup.2/(GT.sup.2)-
). A brain mask generated from the QSM reconstruction can be used
for the NRMSE calculation. The NRMSE number for each image
represents the average of those from all voxels in the entire
volume such as the brain, for example.
The central undersampled data in FIG. 10 represent MR data acquired
with 50% central undersampling. The MR images are mIP/MIP over 8
slices. At the visual level, the Y.sub..theta.1,.theta.2,TE2(y)
images show better quality compared to the
U.sub..theta.1,.theta.2,TE2(y)) images. Specifically, the veins
pointed to by the arrows are more visible in the
Y.sub..theta.1,.theta.2,TE2(y) images than in the
U.sub..theta.1,.theta.2,TE2(y)) images. In fact, the visibility of
the veins in the Y.sub..theta.1,.theta.2,TE2(y) images is similar
to that in the GT images. Also, at the quantitative level, the
NRMSE values are smaller for the Y.sub..theta.1,.theta.2,TE2(y)
images than for the U.sub..theta.1,.theta.2,TE2(y)) images, which
indicates that the Y.sub..theta.1,.theta.2,TE2(y) images are closer
to the GT images than the U.sub..theta.1,.theta.2,TE2(y))
images.
The methods and system described herein provide various techniques
for generating improved images of anatomical regions scanned using
two or more flip angles and two or more echo times. These methods
and systems should not be interpreted as limited to human brain and
can be used for other anatomical regions or organs. Also, while the
figures depict three-dimensional (3D) k-spaces, the imaging
approaches and techniques described herein also apply to
two-dimensional (2D) MR data. Furthermore, the methods and system
described herein may be used to construct other types of MR images
than those disclosed herein. In addition, the imaging system 100 or
the MRI scanner 102 can implement any combination of the methods or
processes described herein.
A person skilled in the art should appreciate that processes
described in this disclosure can be implemented using computer code
instructions executable by a processor, such as processor 104. The
computer code instructions can be stored on a non-transitory or
tangible computer-readable medium such as the memory 106. The
memory 106 can be a random access memory (RAM), a read only memory
(ROM), a cache memory, a disc memory, any other memory, or any
other computer readable medium. Processes described in this
disclosure can be implemented by an apparatus including at least
one processor and/or memory storing executable code instructions.
The code instructions when executed by the at least one processor
can cause performing any of the processes or operations described
in this disclosure. The apparatus can be, for example, the MRI
scanner 102, a computer device or other electronic device
associated with the MRI scanner 102.
* * * * *